Computer simulation for stability performance of sandwich annular system via adaptive tuned deep learning neural network optimization

Abstract

In this article with the aid of adaptively tuned deep neural network (DNN), dynamic stability analysis of the sandwich structure has been investigated. Due to finding the design-points features, the numerical solution procedure called two-dimensional generalized differential quadrature technique has been applied to the ordinary differential equations of the current structure system acquired on the foundation of the kinematic theory with refined higher order terms. Also, the involved parameters with the optimum values in the fully-connected neural network mechanism are obtained via momentum-based optimizer. For modeling a moderately thick structure, higher order terms of shear deformation are chosen. For stability analysis of the current structure the design points considering the method of adaptive learning is presented. For analysis of the current structure 'accuracy (used for determining the design-points) is presented through than the published outcomes in the literature. The outcomes of accuracy section of the current research show that the DNN-based model in analysis of the sandwich structure has less error than other models. The results show that the current momentum-based optimizer can be good tool for future researches about stability analysis of the various structure due to its good accuracy.